THERMODYNAMICS OF ATMOSPHERE VON BEZOLD 275 



From these remarks on the cirrus overflow we now again turn our 

 attention to the thunderstorm cumulus itself. 



When this cumulus penetrates upward into regions where the 

 temperature sinks considerably below freezing the subcooling of 

 the cloudy elerrfent (vapor) must finally attain its limit and some 

 exterior shock will alonebe needed to initiate a sudden freezing. But, 

 as was shown in the first part of this memoir, a warming and sudden 

 increase of pressure must go hand in hand with this freezing. Of 

 course an expansion must follow this rise of pressure and thus we 

 may explain the fact that new cumuli of considerable size often 

 suddenly burst forth from the thunderstorm cumulus. 



Thus, on the 6th June, 1889, on the Summit of the Santis, Assmann 

 took some photographs of an increasing thundercloud, from which 

 in a very short time there broke out of the cloud atypical cumulus 

 dome which afterwards developed into a cirrus of mushroom shape 

 with broad cirrus screen. 



This change of form corresponded completely with the above 

 given formation of the cirrus screen. 



When in consequence of such a dissipation of the subcooled con- 

 dition as above assumed, and as appears to make itself known in 

 the variations of pressure, the now frozen masses are thrown far 

 above the altitude that they would have attained in the subcooled 

 condition, then, after the extinguishment of the upward impulse 

 there must occur a sinking downward, at least of the heavier masses, 

 while the current that blows through the cloud, and is the cause of 

 the whole phenomenon, still continues and thus gives occasion to 

 the formation of the cirrus screen. 29 



According to the above-mentioned investigations of Assmann on 

 the Brocken, as well as in accord with more recent observations by 

 him 30 , such subcooled fog particles never form crystals of ice or 

 snow, but only make lumps of ice without internal structure. 



But sleet is formed by their combination. The assumption of a 

 sudden freezing of subcooled fog particles or very small droplets 

 therefore explains the formation of both sleet and hail without any 

 difficulty. 



At first the frozen subcooled droplets unite into a little pellet of 

 sleet, since they, by falling and striking other subcooled droplets, 

 probably bring these also to freezing and at the same moment con- 

 geal into one mass. When these pellets drop into lower strata in 



29 Compare also Moller: Met. Zeit.. Vol. II, 1890, pp. 220-222. 



30 Met. Zeit., VI, 18S9, pp. 339-342. 



